Chillers - Efficiency

[NickParker]

Which of the following produces maximum cooling efficiency?
1) 2 No's of chiller system with low capacity
or
2) 1 No of chiller with larger capacity

 

[MFJewell]

Depends on your load profile.

 

[MFJewell]

I'll also add, at my facility we have 27 chillers total (range from 1330 ton to 8000 ton) for a total cooling capacity of 112,660 tons of cooling as well as a TES tank equivalent to one 8000 ton chiller for 8 hours. Our load swings drastically throughout the year.

 

[MedicineEng]

Nick:
Your question is so vague that you will not be able to obtain a meaningful answer.

 

[Mech110]

I think the answer lies in IPLV and Full Load Efficiencies of the chillers. As far as my limited knowledge is concerned, chillers are maximum efficient at part loads. Therefore using multiple chillers designed at part loads not only gives redundancy but also greater efficiency though at the cost of greater initial investment.

 

[Bos88]

May be this will help. I ran an example comparing 1 x 800 TR vs 2 x 400 TR using chillers with and without VFD. The chillers were all selected with same Full load efficiency and NPLV values (VFD or non-VFD). I also selected the pumps and cooling towers. The overall installed efficiency was also identical for each case. The design conditions were set at 58F-42F chilled water and 85F-95F condenser. Design wet bulb of 78F was used to select the cooling tower.

Case1: building load is 600 TR and ambient wet bulb has now reduced to 65F. In this scenario we have 1 chiller (800R) running at 75% load vs 2 Chillers (400 TR) also running each at 75%. With constant speed chiller, the overall kW was the lowest when running only 1 chiller (800 TR). With VFD the overall consumption was in favor for the 2 x400 TR chillers. I ran the analysis with different cooling tower fan speed (100%, 70% and 50%) and the resulting condenser water temperatures, and the conclusion remains the same.

Case2: Building load is 300 TR and ambient wet bulb was kept at 65F. In this scenario, we have 1 chiller (800TR) operating at 37.5% vs 1 smaller Chiller (400 TR) operating at 300 TR (75% load). The overall kW consumption was the lowest for the smaller chiller with or without VFD. I also looked at different cooling tower fans speed and it was always in favor of the smaller chiller. Simply here the large chiller is operated at low load (37.5%) which is not a sweet spot for efficiency.

As mentioned by others, the load profile is important. If your load profile has a lot of hours with low load, better have 2 smaller chillers. if your load is fairly constant, then 1 large chiller will work well although you might want to have some redundancy and will most likely look at having multiple chillers. No definite answer on your question. you probably need to run comparison like I did based and look at total system consumption (Chillers, towers, pumps) to see what is best for your application.

If your question was: is it better to run 2 chillers at part-load vs 1 Chiller at higher load in existing facility? then the answer is 1 chiller. For example, let's say you have 2 x 700 TR chillers (1400TR building load). The load is now 600 TR. Do you run 1 chiller at 85.7% capacity or 2 chillers at 300 TR (around 43% load each)? I ran the performances with different tower fan speed for chillers with and without VFD, and running one chiller system was always better. The efficiency of operating 2 chillers at part-load with VFD was actually better than 1 chiller at higher load (also with VFD) but when you add pumps and towers, the overall kW consumption turns out to be higher for the 2 chillers.

In this case, the best option I found is to operate 1 chiller and spread the condenser flow over 2 cooling towers (each tower sees 50% water flow). Set the fan at 70% speed which saves power and increase as well the liquid to gas ratio resulting in colder water temperature. This improves the chiller efficiency and is beneficial for chillers with or without VFD.

Hope this helps

 

[Bos88]

May be this will help. I ran an example comparing 1 x 800 TR vs 2 x 400 TR using chillers with and without VFD. The chillers were all selected with same Full load efficiency and NPLV values (VFD or non-VFD). I also selected the pumps and cooling towers. The overall installed efficiency was also identical for each case. The design conditions were set at 58F-42F chilled water and 85F-95F condenser. Design wet bulb of 78F was used to select the cooling tower.

Case1: building load is 600 TR and ambient wet bulb has now reduced to 65F. In this scenario we have 1 chiller (800R) running at 75% load vs 2 Chillers (400 TR) also running each at 75%. With constant speed chiller, the overall kW was the lowest when running only 1 chiller (800 TR). With VFD the overall consumption was in favor for the 2 x400 TR chillers. I ran the analysis with different cooling tower fan speed (100%, 70% and 50%) and the resulting condenser water temperatures, and the conclusion remains the same.

Case2: Building load is 300 TR and ambient wet bulb was kept at 65F. In this scenario, we have 1 chiller (800TR) operating at 37.5% vs 1 smaller Chiller (400 TR) operating at 300 TR (75% load). The overall kW consumption was the lowest for the smaller chiller with or without VFD. I also looked at different cooling tower fans speed and it was always in favor of the smaller chiller. Simply here the large chiller is operated at low load (37.5%) which is not a sweet spot for efficiency.

As mentioned by others, the load profile is important. If your load profile has a lot of hours with low load, better have 2 smaller chillers. if your load is fairly constant, then 1 large chiller will work well although you might want to have some redundancy and will most likely look at having multiple chillers. No definite answer on your question. you probably need to run comparison like I did based and look at total system consumption (Chillers, towers, pumps) to see what is best for your application.

If your question was: is it better to run 2 chillers at part-load vs 1 Chiller at higher load in existing facility? then the answer is 1 chiller. For example, let's say you have 2 x 700 TR chillers (1400TR building load). The load is now 600 TR. Do you run 1 chiller at 85.7% capacity or 2 chillers at 300 TR (around 43% load each)? I ran the performances with different tower fan speed for chillers with and without VFD, and running one chiller system was always better. The efficiency of operating 2 chillers at part-load with VFD was actually better than 1 chiller at higher load (also with VFD) but when you add pumps and towers, the overall kW consumption turns out to be higher for the 2 chillers.

In this case, the best option I found is to operate 1 chiller and spread the condenser flow over 2 cooling towers (each tower sees 50% water flow). Set the fan at 70% speed which saves power and increase as well the liquid to gas ratio resulting in colder water temperature. This improves the chiller efficiency and is beneficial for chillers with or without VFD.

Hope this helps

 

[Bos88]

Sorry somehow I posted twice my answer

 

[nglty]

Bos88
What software did you use for such modeling and we would appreciate to know you modeled the scenario.

Back to your post. Quiet a few good points. I may add the following:
1. You can monitor your chiller plant amp draw through trending and determine at which loads two chillers use more energy than one chiller, Your BAS can be programmed to switch from two to one chiller when amp reading is more advantageous, meeting the load of course.
2. The other thing about trending and tracking data, one needs to look at the entire plant KW usage, i.e. including tower fans and pumps. Some DDC controls manufacturers claim to have such algorithms tracking entire plant KW.
3. Be cautious of the condenser water "As-cold-as-possible" being best for chillers, not true, studies have shown that at part load, anything below 65F, the chiller use more energy, at full load, anything below 70F, you start using more energy. Also, the 85/95F is a myth, nowadays, we read studies that indicates a 14F degree range is better for cooling towers than the 10F
4. Google Condenser water reset and you will find some interesting things.

Good on y'all.

 

[Bos88]

I did not use any modeling tool. I actually ran selections for chillers, towers and pumps. Got the condenser water data at different fan speed using the manufacturer selection program and the part-load data from the chiller manufacturer program as well

I used 85F-95F for the sake of the evaluation. There are 1000's possibilities on how to evaluate the overall plant efficiency. I simply wanted to provide some kind of answer to the question from NickParker, so I used standard values

With regards to the Condenser Delta T, I fully agree but I would not qualify 14F as the optimum temp either. I don't think it is an absolute rule either. Depends of system size, design wet bulb, cooling tower selection, distance between tower and chiller. I have done a lot of those evaluations as part of my job and sometimes 12F is better than 14F and sometimes it is the opposite based on what the client wants: first cost, operating costs or a balance between both

Agree with the "as cold as possible statement" but this also changes a bit based on VFD vs Non-VFD chillers, single compressor or multi compressors chillers, single stage vs multistage, screw compressors vs Traditional centrifugal Vs mag bearing. No right or wrong answers. My philosophy is always looking at system efficiency on case by case.

In my comparison, I played with the cooling water flow and air flow by sequencing more towers to operate with lower flow rate. I compared running 1 chiller at 600 TR with 1 tower on vs 2 towers (at 50% of their design water flow) each with fan speed at 70%. If I run 1 tower even with fan speed at 100% (40 BHP), the tower water is 73.2F at 65F WB. By running 2 towers each with 50% water flow with 70% fan speed, water drops to 69.8F (about 4 deg F lower) and the fans are now at 13.7 BHP each. So this saved on tower KW and chiller kW. Overall, applying this simple logic saved 23 kW at 600 TR with non-VFD chiller and 32 KW with VFD. Not much savings (could be significant in $ over the course of the year) and this is a simple control logic to implement which works quite well with different wet bulb. Always found operating N+1 towers (within their flow limit) and keeping fan speed at 60-70%.provides savings in operation.

 

[nglty]

Thanks for the reply Bos88 and thank you for elaborating.

Quick question: How do you get the Cooling tower flow to go to 50%? VFD's on CW pumps? these are typically constant speed operation, as there is not much means to throttle the flow. Unless you bring more towers (more fans) whiled maintaining one pump.
I see some VFD's on condenser water pumps but when you look at their operation, they are always around 58Hz. Not much pump savings from VFD's.

 

[lukaiENG]

@BOS88 I'm not sure about your location but be careful about reducing flow across cooling towers in extreme cold conditions. If the fill pack is not fully wetted you can run into icing issues.
This doesn't preclude you from reducing flow at 'low ambient conditions', just make sure you have an 'extreme low ambient condition' sequence in your controls.

 

[Bos88]

nglty, yes I bring 2 towers and 1 pump.
Most likely the reason you see VFD on condenser pump is to balance the flow during commissioning. The VFD gives some flexibility for this. Once they have the desired flow, they often run the pump at constant speed.

 

[Bos88]

lukaiENG,
agree with you. You need as well to check with cooling tower supplier to make sure flow reduction can be done (newer tower models can get to 50% and some to 30% but older models might be limited to 70% max). This also changes if is a crossflow or counterflow (the counterflow nozzles need to be selected according to the flow variation you want to do). SPX marley has a paper on it on their web site.